VividKit Crystal clear AI coding
VI
Choose Your Setup Get Started with AgentKit Continue with ClaudeKit Migrate from ClaudeKit to AgentKit Clean cutover from ClaudeKit What Is AgentKit? How AgentKit Works AgentKit Kits, Skills, Agents & Hooks AgentKit Runtime Support AgentKit Desktop App AgentKit Configuration AgentKit CLI Quick Reference AgentKit Skills Cheatsheet AgentKit Workflows AgentKit Statusline Update AgentKit CLI & Kits AgentKit Helper Clean reinstall of AgentKit AgentKit Troubleshooting What is ClaudeKit? ClaudeKit CLI Guide Command Finder Commands Cheat Sheet Workflow Recipes CLI Quick Reference Claude Mechanics Permission Modes ClaudeKit Hooks Coexistence ClaudeKit x Codex CK → Other coding tools How ClaudeKit Works Inside ClaudeKit Finding Your Unknowns Session Recovery Fix from Logs UI/UX Guide Accounts Switcher IDE Config Remote Control Codex App System Diagramming UI Review Gate Promotions Support VividKit

Quick Ref / Command

Engineer Kit

/ak:ak

Safe AgentKit CLI operation

Guides safe use of the ak control-plane CLI for init, kit, skills, plan, journal, doctor, recover, self-update, login, scope decisions, lifecycle mutations, and JSON output interpretation.

01

Triage intent

02

Inspect help

03

Use parseable reads

04

Confirm scope

Rule 01

ak:ak runs the ak binary safely; ak:agentkit decides which skill should run.

Rule 02

Authoritative order is ak <cmd> --help, then ak <cmd> --json, then generated command-reference.

Rule 03

Read-only inspection precedes lifecycle mutation.

Rule 04

Project scope and user/global scope must never be blurred.

Execution Map

Core lanes from input to output, aligned with the command's real execution path.

01

Start

Input, scope, route

  1. 1 Triage intent Classify the requested ak action as read-only, mutating, or diagnostic before selecting a command family.
  2. 2 Inspect help Run ak <cmd> --help for the installed binary; command-reference is only a starting index that may lag.
02

Work

Agent / skill execution

  1. 3 Use parseable reads For read-only scripted work, prefer --json --no-interactive and parse schema_version, kind, and data instead of scraping prose.
  2. 4 Confirm scope Decide whether the command acts on project state, user/global adapter homes, kit install source, or runtime delivery mode.
03

Verify

Gate, review, validation

  1. 5 Prefer status first Before update, refresh, uninstall, recover, or self-update, run the corresponding doctor, validate, list, verify, backups, versions, or changelog path.
  2. 6 Protect state Preserve unknown files, avoid destructive resets, snapshot before recovery, and run smoke tests only in temp homes with all adapter paths redirected.
04

Close

Report, handoff, artifact

  1. 7 Report exactly Return the exact command, resolved scope, JSON envelope when relevant, disk changes, conflicts skipped, and unresolved constraints.

Syntax, arguments, subcommands

How to invoke the skill: syntax, positional arguments, shared options, then each subcommand with its own syntax and outcome.

Syntax

/ak:ak [goal or subcommand]

Arguments

[goal or subcommand] Required

Goal or ak subcommand

Natural-language request or exact ak subcommand to run, inspect, or interpret. Include the intended project or user/global scope when relevant; this does not author skills or route generic work.

/ak:ak "inspect ak kit list-kits as a read-only JSON command"

Don't skip the required step

Each item contrasts a tempting shortcut with the required approach.

Tempting shortcut

The generated reference listed the flag, so it is safe.

Do this instead

The installed binary help is current; references may lag after self-update.

Tempting shortcut

Use --yes to make automation smoother.

Do this instead

On mutating commands, suppressing prompts needs explicit user approval.

Tempting shortcut

Recovery can be rerun if it fails.

Do this instead

Recovery is unsafe without a fresh backup and confirmed scope.

Sample Prompt

Concrete invocations for each flag, subcommand, mode, or route available in the live workflow.

Inspect installed kits Recommended 
/ak:ak "run ak kit list-kits as a read-only JSON inspection"
Use when:
Use when the next action is inventorying installed kits or interpreting parseable ak output.
Expected:
Inspects ak kit help, classifies list-kits as read-only, uses --json --no-interactive when scripted, and explains schema_version, kind, and data.
Prepare a kit refresh 
/ak:ak "prepare ak kit refresh for engineer after validating scope"
Use when:
Use before a mutating kit lifecycle command so scope and read-only checks are explicit.
Expected:
Runs the read-only doctor/list-kits/validate path first, resolves project versus user/global scope, and keeps confirmations unless the user explicitly approves suppression.
Recover safely 
/ak:ak "recover this project AgentKit install safely"
Use when:
Use when recovery or backup restoration is requested and backup/scope discipline is mandatory.
Expected:
Checks recover help, confirms the target scope, verifies or creates a current backup before mutation, previews conflicts, and reports exactly what changed.

Handled Scope

  • Kit lifecycle
  • Skill inventory
  • Plan and journal CLI
  • Diagnostics and recovery
  • Scope and adapter safety

Next